organic reaction mechanisms 1965€¦ · organic reaction mechanism generally and fairly uniformly,...

ORGANIC REACTION MECHANISMS 1965 An annual survey covering the literature dated December 1964 through November 1965

B. CAPON University of Leicester

M. J. PERKTNS King’s College, University of London

C. W. REES University of Leicester

I N T E R S C I E N C E P U B L I S H E R S a division of

John Wiley & Sons London - New York - Sydney

ORGANIC REACTION MECHANISMS - 1965

ORGANIC REACTION MECHANISMS 1965 An annual survey covering the literature dated December 1964 through November 1965

B. CAPON University of Leicester

M. J. PERKTNS King’s College, University of London

C. W. REES University of Leicester

I N T E R S C I E N C E P U B L I S H E R S a division of

John Wiley & Sons London - New York - Sydney

Copyright 6 1966 by John Wiley & Sons Ltd. All rights raKrved Library of Congress Catalog (3rd Number 66-23143

Made and printed in Great Btitain by William Clowes and Sons Limited, London and Beccles

Preface

This book is a survey of the work on organic reaction mechanisms published in 1966.* For convenience, the literature dated from December 1964 to November 1966, inclusive, was actually covered. The principal aim has been to scan all the chemical literature and to summarize the progress of work on organic reaction mechanism generally and fairly uniformly, and not just on selected topics. Therefore, certain of the sections are somewhat fragmentary and all are concise. Of the 2000 or so papers which have been reported, those which seemed at the time to be the more sqyficant are normally described and discussed, and the remainder are listed.

Our other major aim, second only to comprehensive coverage, has been early publication since we felt that the immediate value of such a survey as this, that of “current awareness”, would diminish rapidly with time. In this we have been fortunate to have the expert cooperation of the London ofice of John Wiley and Sons.

If this book proves to be generally useful, we will continue these annual surveys, and then hope that the series will have some lasting value; some form of cumulative reporting or indexing may even be desirable.

It is not easy to deal rigidly and comprehensively with so ubiquitous and fundamental a subject as reaction mechanism. Any subdivision ia a necessary encumbrance and our system, exemplified by the chapter headings, has been supplemented by cross-references and by the form of the subject index. We should welcome suggestions for improvements in future volumes.

February 1966 B.C. M.J.P. C.W.R.

* In Chapter 1, only, an acoount of earlier work is given in some detail, since an introduction to the current oontrovemy on claesioal and non-classical carbonium iona eeemed particularly timely.

Con tents

1. Classical and Non-classical Carbonium Ions . . 1 Introduction . . 1 Bicyclic Systems . . 13 Phenonium Ions . . 31 Participation by Double and Triple Bonds . . 37 Cyclopropyl Carbonium Ions . . 43 Cyclopropyl Cations , . 44 Other Stable Carbonium Ions and Their Reactions . . 45

2. Nucleophilic Aliphatic Substitution . . 50 Ion-pair Return and Related Phenomena . . 50

Isotope Effects . . 62 Deaminations . . 63 Fragmentation Reactions . . 66 Displacement Reactions a t Elements other than Carbon . . 68 Ambident Nucleophiles . . 72 Other Reactions . . 75

3. Electrophilic Aliphatic Substitution . . 81 4. Elimination Reactions . . 90 5. Addition Reactions . . 104

Electrophilic Additions . . 104 Addition of halogens and nitrosyl halides . . 104 Hydration and related additions . . 108

Nucleophilic Additions . . 113 Radical Additions . . 114 Diels-Alder Reactions . . 123 Other Cycloaddition Reactions. . . 127

6. Nucleophilic Aromatic Substitution . . 133 The Bimolecular Mechanism . . 133 Meisenheimer and Related Complexes . . 137 Substitution in Polyfluoro Aromatics . . 139 Heterocyclic Systems . . 141 Diazonium Decomposition . . 143 Other Reactions . . 144 Benzyne and Related Intermediates. . . 147

Neighbouring-group Participation . . 55

Epoxidations . . 111

viii contents

7. Mad and Elwtrophilio Arormrtio Subtitution . Radical Substitution . Electrophilic Substitution .

8. Moleanlarb-gements . Aromatic Rearrangements . Cope and ReIated Rearrangements; Valence-bond Isomerieation . Radical Rearrangements . . Other Rearrangements . Radical-forming Reactions . Rwtions of Free Radicale .

9. RaaidIleratiolM .

Oxygen radicals . Radical abstraction and displacement prmmee Radical couphg and disproportionation . Miscellaneous data on free radioals

Oxygen radicals . Nitroxides Nitrogen radicals. . Carbonradicala . Radicalanions . Radical cations .

10. Csrbenes and Nitrenes . 11. Reactions of Aldehydes and Ketonem and Their Derivatives .

. .

Stable Radicals and Electron Bpin Resonance Studies .

Formation and Reactions of Ace* and Ketsls . Reactions with Nitrogen Bases. . Enolization and Relsted Reactions . Other Reactions ,

12. Reactions of Acids and Their Derivatives . Carboxylio Acids . Non-carboxylic Acids .

18. Phofochemktry . 14. Oxidatione and Reduction8 .

Oxidation of Olefine . Oxidation of Hydroxyl and Carbonyl Compounds Other Reactions .

.

. Author Index ,

Subjeot Index .

164 164 169 172 172 179 184 187 193 193 201 201 202 209 21 1 214 214 216 217 218 219 221 222 237 237 242 246 266 260 260 280 286 301 301 304 307 316 346

CHAPTER 1

Classical and Non-classical Carbonium Ions

Introduction

The questions which a t present stimulate most interest in this field are whether or not certain carbonium ions have bridged or non-classical structures and whether this bridging is developed in the transition states of reactions in which these ions are thought to be intermediates.l One of the earliest proposals that a carbonium ion does not have a classical KekulB structure was that of Nevell, de Salas, and Wilson2 for the ion from camphene hydrochloride (1). In aprotic solvents this compound undergoes a stereospecific rearrangement to yield isobornyl chloride (2) and not the more stable bornyl chloride (3). It was therefore suggested that, rather than there being two

intermediate ions (4) and (5) with classical structures, there was a single ion, mesomeric between them. This ion can then be written as (6)3 and it has

become customary to ascribe to the split bond, or bridge, the property of preventing attack from the endo-direction. This idea was developed by Hughes, Ingold, and their colleagues, who suggested that the 600-fold greater

l For reviews on the subject of carbonium ions and their reactions see D. Bethell and V. Gold, Quart. Rev. (London), 12, 173 (1958); N. C. Deno, Progr. Phys. Org. Chem., 2. 129 (1964); Chem. Eng. News, 6th Oct., 42, No. 40, p. 88 (1964). T. P. Nevell, E. de Selas, and C. L. Wilson, J . Chem. Soc., 1030, 1188. Suggestion of C. K. Ingold quoted by H. B. Watson, Ann. Rep. Progr. Chem. (Chem. Soc. London), 36, 197 (1939). 1+

2 0rgan;C Reaction Mechcrnisms 1965

rate of solvolysis of camphene hydrochloride compared with t-butyl chloride and the 105-fold greater rate of solvolysia of iaobornyl chloride compared with bornyl chloride and pinacolyl chloride could be explained if this bridging were developed in the transition states (7) and (8) of these reactions.* They

(7) (8)

referred to these bridged ions as “synartetic” iona and to the rate enhancements thought to be caused by this bridging in a transition state as “spartetic acceleration”. The spartetic ion (6) was discussed as a resonance hybrid between the two classical ions (4) and (5), despite there being sub- etantial differences in the positions of some of the nuclei in the two structures.

The concept of bridged or non-classical ions was also taken up by Winstein and Trifans to explain why em-norbornyl p-bromobenzenesulphonate (9; Bs = p-BrCeH,S02) undergoes acetolysis 360 times faster than its endo- isomer (10) and yields, as the product of substitution, almost exclusively a-norbornyl acetate.” This striking result, in which the normal tendency of

carbonium ion reactions to yield racemic or predominantly inverted products is completely reversed, led Winstein and Trifan to suggest that the reaction involved a bridged or non-classical ion (11). The greater rate of solvolysis of the em-isomer was explained by supposing that this bridging was developed in the transition state: i.e. that the reaction involved partici-

4F. Brown, E. D. Hughes, C. K. Ingold, and J. F. Smith, Nature, 168, 66 (1951); C. K. Ingold, “Structure end Mecbniam in Organic Chemiatry,” Cornell University Preee, Itheoa, N.Y., 1963, pp. 514-4623. 6S. Winatein and D. 9, Trifan, J . Am. Chm. he., 71,2963 (1949); 74,1147,1154 (1953). eThe moat reoent 5gure ia that aoetolyrris of em-norbornyl toluene-p-sulphomte yields

0.05 f 0.02% of endo-acwtate at 100’ and OB. 0.01% at SOD [H. L. Qoering and C. B. Sohewene, J . Am. Chm. Xoc., 87, 3516 (1965)l. It b alao been reported tbt solvolysia of the em-pbromobenzeneaulphonate in 76% equeoue acetone yielda leaa than 0.02% of d - n o r b o m o l [S. W W i , E. Clippinger, R. Howe, and E. Vogelfmger, J . Am. Chem. Sm., 87, 376 (1965)l.

Classical and N o n - c h s h l Carbonium Ions S

pation by the C,l,-C,,, bonding electrons. Winstein refers to the rate enhancement brought about in this way as “ anchimeric assistance”.’

Supporting evidence for this proposal was provided by the observation that the exo-norbornyl acetate obtained from optically active m-norbornyl p-bromobenzenesulphonate was, within experimental error, completely racemic.6*8 This is, of course, to be expected if ion (11) is an intermediate since it has a plane of symmetry. It was also observed that the rate of racemization was about 3.5 times greater than the rate of formation of titratable acid. This means that the exo-norbornyl p-bromobenzenesulphonate is itself undergoing racemization and it was suggested that this involves dissociation to, and recombination of, an ion pair (12)) the whole process being referred to as “ion-pair return”. Thus only 29% of the acetate in the product comes directly from optically active starting material, but this also was shown to be racemic within experimental error.

That the processes which occur are even more complex than this was shown by Roberts, Lee, and Saundersg who studied the acetolysis of exo- norbornyl p-bromobenzenesulphonate labelled specifically in the 2- and the

3-position with carbon-14. Instead of obtaining exo-acetate labelled only in the 1-, 2-, 3-, and 7-positions (as 13) which would be expected if there was just one intermediate bridged ion they found that the label was in positions 1,2 , 3, 5, 6, and 7 as (14). This additional scrambling of the label must result from a C,,,+C,,, hydride ion shift to give a second ion (15). The postulated

8. Winstein, C. R. Lindegren, H. Marshall, and L. L. Ingraham, J. Am. Chem. Soc., 76, 147 (1963).

* The latest value is that the survival of optical activity on aolvolysia of optioally active e~o-~-bmmobenzenesulphonate (9) in acetic acid, 80% dioxan, or 76% acetone at 26’. and 90% dioxan at 60”, ia less than 0.06% [S. Winstein, E. Clippinger, R. Howe, and E. Vogelfanger, J. A.m. Chem. Soc., 87, 376 (1966)l.

@ J. D. Roberts and C. C. Lee, J. Am. Chem. Soc., 78, 6009 (1961); J. D. Roberts, C . C. Lee, and W. H. Saunders, ibid., 76, 4601 (1964).

4 Organic Reactima Mmhfibms 1965

non-classical ion therefore appears to have the, at first sight, unusual property of being attacked by external nucleophiles only in the exo-direction, but by the migrating hydride ion in the enclo-direction. The migration was, however, rationalized as involving an intermediate face-protonated, or edge- protonated, (17),6 cyclopropane, the former being also referred to as a nortricyclonium ion.

At about the same time another striking stereochemical result was obtained by Crsm,lo who showed that the product of substitution from the acetolysis of optically active L-them-ap-dimethylphenethyl toluene-p-sulphonate (18) contained 96% of racemic them-acetate, and 4% of erythro-acetate. The total yield of substitution products waa 53%; 35% of olehs were also obtained. The strong preference for retention of contigumtion was explained as resulting from the intervention of a bridged phenonium ion (U), and this accounta also for the themacetate's being almost completely racemic since this ion has a plane of symmetry. The acetates obtained from the D-f?ythro-ester (20) consisted of 94% mmjtheo- and 5% D-themacetate, obtained in total yield of 68% along with 23% of olehs. This result was explained as being due to the intervention of bridged ion (21) which, since it is asymmetric, should yield active products. Ion-pair return was also found on acetolysis of the L-three-toluene-p-sulphonate for which the rate of racemization is five times as great as the rate of formation of titratable acid. This means that only 20% of the product could have been formed directly from optically active toluene-p-sulphonate. It is not possible to study ion-pair return in the acetolysis of the erythro-toluene-p-sulphonate by polarimetric measure- ments becauae this doe0 not result in racemization; but ion-pair return has recently been detected and ita rate measured by a carbon-14 labelling technique.ll

Cram has obtained similar results for the formolyses of these compounds, and in these reactions the stereospecificity is even greater.1°

The anchimeric assistance that can be associated with participation by the phenyl groups in these systems is small. Thus the rates of the acetolyses

10 D. J. Cram, J. Am. Ckm. Soc., 71,3863 (1949); 74,2129,2137 (1962); for a recent snmmary of thie work nee D. J. Cram, J . Am. Claem. Boc., 88,3767 (1964).

l1 W. B. Smith end M. Showdter, J. Am. C h . 8m., 86,4136 (1964).

Classical awl Non-classical Carbonium Ioris

of both the up-dimethylphenethyl toluene-p-sulphonates are less than that of s-butyl toluene-p-sulphonate; l2 but when allowance is made for the electron-withdrawing inductive effect of the phenyl group, and when the rate of ionization (as determined polarimetrically) of the threo-toluene-p- sulphonate rather than its rate of acetolysis is used, the rate enhancement due to phenyl participation is computed to be 24-fold.1° The validity of this procedure is uncertain since there is the possibiIity that ionization to, and recombination from, an ion pair occurs without racemization, and since also the extent of ion-pair return occurring in the acetolysis of s-butyl toluene-p-sulphonate is unknown.

Another class of compound in whose solvolyses the intervention of non- classical ions has been postulated are derivatives of cyclopropylmethyl alcohol. The evidence for this is that the solvolyses of these compounds frequently yield cyclobutyl and 3-butenyl derivatives. Thus the acetolysis of cyclopropylmethyl chloride yields cyclopropylmethyl and cyclobutyl acetate in the ratio 2.6: 1, a small amount of 3-butenyl acetate, and a 1.7: 1 mixture of cyclobutyl and 3-butenyl ch10ride.l~ Acetolysis of cyclobutyl toluene-p- sulphonate l4 and formolysis of 3-butenyl toluene-p-sulphonate l5 yield similar mixtures. An experiment with specifically deuterated cyclopropylmethyl chloride indicated considerable skeletal rearrangement in the cyclopropylmethyl chloride isolated from a partly solvolysed reaction mixture.l8 The solvolyses of the cyclopropylmethyl compounds also proceed at enhanced rates,13J7 ethanolysis of the benzenesulphonate, for instance, being 500 times

l2 S. Winstein, B. K. hlorse, E. Grunwald, K. C. Schreiber, and J. Come, J. Am. Chem. SOC.,

l3 J. D. Roberts and R. H. Mazur, J . Am. Chem. Soc., 78,2609 (1961). 14 J. D. Roberts and V. C. Chambers, J . Am. Chem. Soc., 78, 6034 (1961). l5 K. L. Servia and J. D. Roberts, J. Am. Chem. SOC., 86,3773 (1964). 1eM. C. Caserio, W. H. Graham, and J. D. Roberta, TetraWrm, 11, 171 (1960). 1' C. G. Bergstrom and S. Siegel, J . Am. Chem. Soc., 74, 146 (1962).

74, 1113 (1962).

6 Organic Readion Mechanisms 1!%5

faster than that of ethyl benzenesulphonate. These resulta suggest that ionization of the cyclopropylmethyl, cyclobutyl, and 3-butenyl compounds yield the same ion, or readily interconvertible ions which may either react with solvent or re-form chloride or toluene-p-sulphonate by ion-pair return. Roberts and his co-workers favoured the intervention of an equilibrating set of bicyclobutonium ions (22) to (24) rather than a single tricyclobutonium

ion (25) because in certain non-solvolytic reactions there was incomplete equilibration of the methylene groups.

Non-classical ions have ah0 been postulated as intervening in reactions in which there is participation by double bonds.18 Thus emnorbornenyl p- bromobenzenesulphonate (26) undergoes acetolysis 8000 times more rapidly than its endo-isomerlO and yields about 80% of 3-acetoxynortricyclene (29) and some em-norbornenyl acetate (80).ao*a1 The reaction may therefore

be formulated as involving a “non-classical transition” state (W), leading to a non-classical ion (28). Acetolyais of em-norbornenyl p-bromobenzenesulphonate labelled at C,,, and C,,, with carbon-1 3 yields ezo-acetate which

For reviews eee B. Capon, Quart Rev. (Ldun) , 18,97 (1964); B. Capon and C. W. Reea, Ann. Re@. P r w . Chem. ( O h . Sm. London), 81,231 (1964).

l9 Unpnblbhed work of H. J. Sohmid and I(. C. Sohwiber, reported by 5. Winetein and Ed. Shataveky, J. Am. Chem. Sm., 78,696 (1966). 8. Winste.ii, H. M. Walborsky, and K. Sohreiber, J . Am. Chem. Soc., 72,6795 (1960).

Dl J. D. Roberte, W. Bennett, and R. Armstrong, J . Am. C h . Sm., 78,3329 (lQa0).

Classical and Non-classical Carbonium Ions 7

has lost about one-third of its label a t these positions.22 It thus became necessary to propose that ion (28) underwent rearrangement to its enantio- morph (31) or to a symmetrical ion (32).2a

A much greater rate enhancement is observed in the acetolysis of anti- 7-norbornenyl toluene-p-sulphonate (yielding anti-7-norbornenyl acetate) which proceeds 10" times faster than that of the analogous saturated compound, 7-norbornyl toluene-p-~ulphonate.~~ The r-electron cloud of the double bond of this compound ie particularly well placed to interact with the

Scheme 1

developing carbonium ion at position 7 and the solvolysis has therefore been formulated as shown in Scheme 1.7-Norbornadienyl derivatives under- go solvolysis even more readily; thus the chloride (33) reacts in aqueous acetone at a rate 760 times greater than that of anti-7-norbornenyl chloride and it was suggested that the reaction involved an ion of structure (84) or (35) as an intermediate.24

22 J. D. Roberts, C. C. Lee, and W. H. Saunders, J . Am. Chem. Sm., 77.3034 (1966). a3S. Winstein, M. Shatavsky, C. Norton, and R. B. Woodward, J . Am. C k . ~Yoc., 77,

4183 (1966); S. Winetein and M. Shatavsky, ibid., 78. 692 (1966). 5. Winstein and C. Ordronneau, J . Am. Chem. Soc., 82,2084 (1980).

8 manic R a d o n Mechanisms 1965

This then summarizes the salient evidence available at the end of the 1960’s for the exietence of some of the more important non-classical ions and for their being formed and reacting by way of non-classical transition states. The concept had received almost universal acceptance among organic chemists, so much so that in the last yeam of that decade and in the early 1960’s a large number of non-classical ions were postulated as reaction intermediates, frequently with little supporting evidence. Since about 1960, however, the view has been taken by H. 0. Brown that; in many and possibly all of these reactions, it ie unnecessary to postulate these non-classical or bridged ions and that it is possible to explain the stereochemical and some- times the kinetic resulta without involving non-classical ions or transition states.a6 He has paid particular attention to em-norbornyl systems and made

a detailed study of the effects of substituents in the 1- and 2-positions on the rates and products.2B-a8 On ionization a 1- or Zsubetituted em-nosbornyl derivative would give the same non-classical ion, as illustrated. Hence, if we assume that there is little difference in the free energies of the initial state, which ia reasonable, then in the extreme case in which delocalization of the 1,6-bond is complete in the transition state the effect of a 1- and EL 2- substituent on the rate should be almost the same, and the greater the difference between the effects of substituents in these two positions the less this delocalization must be. The effect on the rate of introducing phenyl and methyl substituents was therefore investigated and it was found that, whereas 2-phenyl-em-norbornyl chloride underwent ethanolysis a t a rate estimated to be 3.9 x lo7 times greater than that for mo-norbornyl chloride, with the 1-phenyl homer the corresponding rate enhancement was only 3.9.ae Provided then that there is no steric inhibition of resonance between the l-phenyl substituent and the developing cationic centre this result means either that there ie little delocalization of the 1,6-bond in the transition state which is akin to saying that it is “classical”, or that the stabilizing effect of a 2-phenyl substituent on the classical transition state (36) is so much greater than on the non-classical one (37) that reaction proceeds exclusively through the former. Nevertheless, it was also shown that the rate

Cf. H. C. Brown, “The Traneition Steta”, Chem. Soc. 8pmial Publ., No. 16, p. 144 (1962). as H. C. Brown, F. J. Chloupek, end Min-Hon Rei, J . Am. Chem. ~Soc., 86, 1246, 1247, 1248

(1904).

Classical and Non-chseical Carbonium Ions 9

of solvolysis of 2-phenyl-ezo-norbornyl p-nitrobenzoate (38) in 60% ethanol is 140 times greater than that of the endo-isomer (39).

A similar situation is found when considering 2-methylnorbornyl deriva- t i v e ~ . ~ ~ A 2-methyl substituent increases the rate of ethanolysis of exo- norbornyl chloride by a factor of about 6 x lo4 (compare 40 and 41), but

Me

(40) (41) 106k(sec-') 4.72 10-4 30.0

at 25"

an additional 1-methyl substituent has only a 4.6-fold effect on the rate of hydrolysis of 2-methyl-m-norbornyl p-nitrobenzoate in aqueous dioxan (compare 42 and 43).27 This is in fact slightly less than that observed in the

(43) (*a) (4) (a) I@ k (aec-1) 2.2 9.4 0.012 0.057

at 50" X n p-NO,C,H,CO

H. C. Brown end Min-Hon Rei, J . Am. Chem. Soc., 86, 6004 (1964). I*

10 Orgamic Readiolt Mffihn&nur 1965

endo-series (compare 44 and M), indicating that there can be little delocalization of the 1,6-bonding electrons in the transition state for the reaotion of the em-isomer. Nevertheless, the em:endo ratio ie about 180 (compare 42 and 44), and the product of substitution of the --chloride (46) in aqueous dioxan is the em-alcohol (47) derived from the tertiary ion exclusively.aa

Brown has interpreted this large difference between the effects of sub- stituenta in poaitione 1 and 2 as meaning that the tertiary m-norbornyl derivatives react via classical transition states and an equilibrating pair of classicalions. These reactions, however, show high m: endo rate- and product- ratios and hence, if Brown’s view is correct, these criteria lose the validity as testa for non-classical ions and transition states.

Brown explains the high m: end0 rate-ratio as due, not to a large rate for the m-homer, but to R small rate for the endo-isomer.as This is attributed to steric hindrance of departure of the leaving group from the endo-isomer. As this group departe, C(21 starte to become spa-hybridized and planar, and its developing p orbital is directed towards the 6,6-endo-hydsogens (see 48). This position has, however, been criticized by Winstein who prefers t o

interpret Brown’s results as meaning that carbon bridging lags behind C-X ionization in the transition state.ag This would presumably mean that the amount of stabilization due to delocalization in the free ion is very much larger than in the transition state, which on the basis of a rate enhancement of ca. lo3 can be computed to be about 4 kcal. mole-’. Since the 2-norbornyl cation can be observed directly by NMR spectroscopy (see p. 23) it is to be hoped that some measure of its stability will be forthcoming.

ps 8. Winstein, J . Am. Chm. Soc., 87,381 (19136). H. C. Brown end H. M. Bell, J. Am. CLm. Soc., 86,5008 (1964).

Classical and N o l z - c h e k l Carbonium Ions 11

It is interesting that 1-methylnorbornyl toluene-p-sulphonate undergoes acetolysis about 60 times faster than mo-norbornyl tol~ene-p-aulphonate.~~ This must mean that there is participation by the 1,6-bonding electrons in this reaction (i.e. a non-classical transition state), but the driving force for this could be the rearrangement to the more stable tertiary l-methylnorbornyl cation which could be classical. It is perhaps significant that, as outlined above, there was no similar rate enhancement in the solvolysis of the symmetrical 1,2-dimethyl derivatives, for here this driving force would be absent.

Brown’s view that the solvolysis of endo-norbornyl derivatives is unusually slow is in disagreement with Schleyer’s i or relation.^^ This is a relationship between the rates of solvolyses and the carbonyl stretching frequencies of the corresponding ketones, and it includes correction terms for inductive effects and for torsional and non-bonded interaction strain. It predicts correctly the rates of solvolysis of a large number of compounds, but compounds whose reactions are considered to be anchimerically assiated show higher reactivities than are calculated. The observed reactivity of endo- norbornyl toluene-p-sulphonate is close to the calculated value but that of the exo-isomer is larger, suggesting that its reaction is anchimerically assisted. There is, however, always the poasibility that some important steric factor unique to the norbornyl system (e.g. steric hindrance to ionization) has been overlooked.

It has been urged that these kinetic studies tell us nothing about the BtrllC-

tures of the carbonium ions but only about the structure of the transition states,3a and that it is the stereochemistry of the reaction products which provide information about the former. On this ground, therefore, Berson considers that the best evidence for the incursion of a non-classical ion in the norbornane system is that replacement occurs exclusively from the

30 P. von R. Schleyer and D. C. Kleinfelter, 138th Meeting A.C.S., New York, September, 1960, Abstracts, p. 43P; 888 also J. A. Berson in “Molecular Rearrangementa”, P. de Mayo, ed., Interscience, New York, 1963, Part 1, p. 182.

An interesting dichotomy of viewpoint on the relevance of rate differences to the structure of the ion, as distinct from the tramition state for its formation, has grown up. The original workers were frequently unconcerned with this distinction and assumed without comment that the rate enhancements could be discussed in terms of the struoturea of the ions them- selves (cf. ref. 6). More recently the views have been taken (see ref. 81) that “Rate comparisons reflect only differences in activation free energies and therefore provide no information about structure after the rate-determining transition state is reached ’0 and that rate comparisons are the “least cogent” evidence for non-classical ions (see ref. 3%). Alternatively it has been written: “Of the three unusual properties associated with non- classical ions, only one, enhanced reaction rate, can Berm to distinguish between these structural alternatives” (see ref. 31), and “Some of the recent discussions of anchimeric essistance to ionization have sounded es if there were no conneotion between the structure of a carbonium ion and that of the transition state leading to it” (we ref. 100).

31 P. von R. Schleyer, J . Am. C b m . Soc., 86, 1854, 1866 (1964).

i a ~ g a n i o W i m M e c k h 1965

em-direction even when C,,, carries a gm-dimethyl This contrasts with the behaviour of the corresponding ketones towartrds hydride reduction, in which a tendency towards exo-attack changes on the introduction of a 7,7-dimethyl group to a tendency towarde &attack. If it is valid to draw an analogy between this reaction and that of a classical carbonium ion with a nucleophile one would expect a 7,7-dimethylnorbornyl cation to react with substantial endo-attack if it were classical, but this has never been observed. This is a telling piece of evidence for the existence of non-chical carbonium ions, but at the moment it is rather narrowly based and it is to be hoped that more experimental evidence wil l be forthcoming to indicate the relative ease of am- and endo-attack in 7,7-dimethylnorbornyl systems (see also p. 266).

Brown has also questioned Hughes and Ingold’s assignment of the &h rate of solvolysis of camphene hydrochloride (49a) compared with that of t-butyl chloride to synartetio acceleration (see p. 2).34 In fact, the rate of solvolysis of camphene hydrochloride is not appreciably larger than that of the almost analogously substituted cyclopentyl chloride (49b). The high

reactivity is therefore most probably caused by steric acceleration and not by participation of a-bonding electrons.

Although accepting that the solvolyses of anti-7-norbornenyl and ‘I-norbornadienyl derivatives involve participation by the double bond (i.e.

involve non-classical transition states), Brown has questioned whether the ions could not better be represented a8 classical structures (60) and (51).36 Some support for this formulation of the 7-norbornadienyl cation has come

33 J. A. Berson in “Molecular Rearrangements”, P. de Meyo, ed., Intarscience. New York,

O4 H. C. Brown and F. J. Chloupek, J. Am. Chem. Boc., 85,2322 (1983). 36 H. C. Brown and H. M. Bell, J . Am. Chem. am., 85,2324 (1983); w also S. Winstein, A, H,

1983, Part 1, p. 130.

Lewin. and K. C. Pande, ibid., 86,2324 (1963).

CzdpsiCal and Non-classical Carbonhn Ions 13

from the NMX spectrum of its hexaflu~roantimonate.~e The signal from the proton a t position 7 occurs a t -3.5 ppm and this ia at rather a high field for a proton attached to carbon carrying a positive charge (usually -9.5 to -13.5 ppm) even when allowance is made for some drawing off of this charge by the double bond.37 It is, however, at approximately the correct position for the 8-proton of a cyclopropylmethyl cation.

The status of non-classical carbonium ions is thus a matter of considerable controversy, but there have been significant developments in this during the last year and these and other aspects of carbonium ion chemistry wil l now be reviewed.38

Bicyclic S y s t 0 m ~ ~ ~

Thia year interest haa continued in the structure of the norbornyl cation and the transition states of its reactions. In a highly illuminating investigation Goering and Schewene40 have measured the rate of the perchloric acid- catalysed loss of optical activity of ao-norbornyl acetate (52) and of its isomerization to the endo-isomer (68). These reactions are thought to involve reversible protonation of the substrate, followed by heterolysis of the conjugate acid of the acetate:

1 Ion or

R ions

36 P. R. Story, L. C. Snyder, D. C. Douglass, E. W. Anderson, and R. L. Kornegay, J . Am. Chem. floc., 85, 3630 (1963). N. C. Deno, Progr. Phyu. Org. Chem., 2, 248 (1964).

Dewar and A. P. Marohand, Ann. Rev. Phye. Chern., 16,321 (1966).

Interscience, New York, N.Y., 1963, Part 1, p. 111.

38For other recent reviews see: (a) W. Hiickel, J . Prukt. Chem., 28. 27 (1966); (a) M. J. S.

38 For a recent review see J. A. Berson in “Molecular Rearrangements”, P. de Mayo, ed.,

‘0 H. L. Goering and C. B. Schewene, J. Am. Chem. Soc., 87, 3516 (1966).

14 Organic Reaction Mechanism 1965

The former rate is then the rate of ionization of the emwetate (52) and the latter the rate of capture of the ion (or ions) by endo-attack. It was thus estimated that exo-attack predominates to the extent of 99.99% and 99.96% a t 26' and loo", respectively. From these results and the equilibrium con- stant for d conversion a potential diagram (Figure 1) was constructed.

I\ iOM I L! 25.3 f 0.1 EA s. e8.4 f Q4

Y FIQ. 1

It is seen that the energy of activation for capture of the norbornyl cation(s) from the em-direction is 4.4 0.7 kcal mol-l less than from the endo-direction and that this (leas the difference in the initial-state energies) is also the difference in the energy of activation for the ionization of exo- and endo-norbornyl acetate. This work thus illustrates oleclrly that it is not the structure of the carbonium ion which directly controls the stereochemistry of substitution but the difference in the energiea of the transition states for reaction of this ion by em- and endo-attack. In this system though, by the principle of microscopic revereibility, these transition states and this energy difference must be the same as for the ionization of the a ~ e t a t e a . ~ ~ In the solvolyses of arenesulphonates and halides thh symmetry will be distorted, but clearly those factors that control the difference in the rates of the em- and endo-isomers will be the same as those that control the product ratio. If therefore one subscribes to Bereon's that in suoh reactions "kinetics alone give no information on the structure of the cationic intermediate (or intermediates) ", one is forced to the conclusion that neither does the am: endo product ratio.

In principle, however, if classical ions are reaction intermediates it should H. C. Brown and a. L. Tritle, J . Am. Chem. Sm., in the prese. Bee ref. 39, p. 118.

Classical and Non-chsical Carbonium Ions 15

prove possible to trap them before the system becomes symmetrical; one reaction in which a substituted classical norbornyl cation derived from a exo-toluene-p-sulphonate has been trapped is reported by Takeuchi, Oshika, and K ~ g a . ~ ~

These workers investigated the methanolysis of the four 6,G-trimethylene- 2-norbornyl toluene-p-sulphonates (54-57). Compound (MI), with the

toluene-p-sulphonoxy group e m and the trimethylene bridge endo, is converted into compound (54) by ion pair return a t a rate 2.2 times that of its solvolysis. Both compounds yield only the em-methyl ethers (68) and (59) and their ratios at 100% reaction are, within experimental error, the same from the two toluene-p-sulphonates. Most of the product a t 100% reaction has however always come from (54), since (56) is converted into (54); and to determine the direct solvolysis product of (56) it was necessary to examine the product early in the solvolysis. When this was done it became clear that compound (56) was yielding more of the corresponding methyl ether (59) than of its isomer (see Table 1). There cannot therefore be a single product- forming intermediate (i.e. a non-classical ion), and the reactions probably

Table 1. Ratea and produots for the methanolysk of the 6,6-trimethylene- 2-norbornyl toluene-p-sulphonates.

Compound 54 55 56 57

Relative rate of 1.00 0.0986 3.84 0.0246

yo of (69) in product 3.2 1.6 7.3 23.4 methanolysis at 74.8"

involve two classical ions (60) and (61). Nevertheless, cornpounds (54) and (56) react faster than their epimers (55) and (57) which contain an endo- tosyloxy group, the rate ratios being 10.4 and 166, respectively (Table 1). Since the reactions of (54) and (56) involve classical ions, the transition states are presumably also classical; hence this rate difference cannot be ascribed to anchimeric assistance and, as it is much greater when the trimethylene bridge is in the endo-position, Brown's suggestion of steric

43 K. Takeuchi, T. Oehika, and Y. Koga, Bull, Chem. SOC. Japn, 88, 1318 (1966).

16 Organic Reaction Mechunisms 1965

hindrance to ionization (see p. 10) of the endo-isomers (66) and (57) seems to offer a reasonable explanation. It should, however, be noted that the ao:endo rate ratio is rather small, so it is possible that the property of reacting via classical ions is unique to this system. It is to be hoped that other systems will be studied in a similar way.

The solvolysea of 2-norbornyl p-bromobenzene-sulphonates with a one- carbon bridge between positions 6 and 6 (i.e. the tricyclo[3.2.1.0a~4]oc~ne system) have also been inve~tigated.~~**~ Compound (68) is converted into

compound (62) by ion-pair return at a rate (k = 2.67 x 8ec-l at 26") about four times faster than its rate of acetolysis, and all three compounds (62), (a), and (64) are partially converted into 3-nortricyclylmethyl p - bromobenzenesulphonate (66) which undergoes acetolysis much more slowly than any of them. The products isolated at complete reaction from (62), (a), and (64) (Table 2) are, within experimental error, identical; clearly not too much emphasis should be placed on this since they all probably result, in the

44 K. B. Wiberg end Q. R. Weneinger, J . o*g. Ohm., 80,2278 (1966). 46 A. K. Colter and R. C. Mueso, J . o*g. Chm. , 80,2462 (1966).

Classical and Non-cEaasiea1 Carboniunt Ions 17

Table 2. The rates and products of acetolysis of the 6-tricycl0[3.2.l.O~*~]ootanyl p-bromobenzenesulphonates.

Products (yo) Minor

mainly 68 67 acetates

p - Bromobenzenesulphonate 105kZ5 105k,, (sec-I) (sec-l)

eZ0,eZO- (62) 1.36 67.5 f 0.8 32.5 endo,exo- (63) 0.68 68.5 rt 0.1 31.5 ezo,endo- (64) 3.35 67.5 rt 0.1 32.5 endo,endo- (65) 3.36 3- Nortricgcly lmet hyl (66) 0.653 58.2 & 0.5 41.8

main, from (62) which is formed by ion-pair return from (68) and (64). It would be of considerable interest to know if the product ratio from (63) as determined early in the reaction, before there had been appreciable isomerization, is the same as that of the product isolated a t complete reaction. The products from the endo,endo-isomer (65) were not reported but are being investigated.

T

(67) (68) (69) (70)

Schleyer, Donaldson, and Watts have shown that charge delocalization to position 6 in the transition state for the acetolysis of exo-norbornyl toluene-p- sulphonate [i.e. as indicated by resonance between structures (69) and (70) for the ion] must be of minor importance since the introduction of a gem-

Table 3. The rates of acetolysis of norbornyl toluene-p-sulphonates at 74.84'.

2-endo- 6,6-Dimethyl-2- 2-exo- 6,6-Dimethyl-2- Norbornyl endo-norbornyl Norbornyl ezo-norbornyl

10% (sec-I) 5.09 0.514 519 36.5

dimethyl grouping a t this position caused a rate retardation rather than a rate enhancement (Table 3).46 This retardation was attributed to an unfavourable steric interaction between the methyl groups and C(l) and C,, in a non- classical transition state. A similar rate depression was observed in the endo-series and this was reported as the first case of significant steric decel- eration in a unimolecdar solvolysis.

48 P. von R. Sohleyer, M. M. Donddaon, and W. E. Watts, J . Am. Chem. Soc., 87,376 (1966).

18 Organic Redion Mechanisms 1965

The effects of gem-dimethyl groupings a t positions 3, 6, and 7 have been investigated by Winatein and his co-worker~.~~ Ion-pair return wae detected in the aolvolysis of 7,7-dimethyl-~-norbornyl p-bromobenzenesulphonate (71) (a) becaw the kinetics were not of the first order and (b) by isolation

from the reaction mixture of the more slowly reacting 3,3-dimethyl-m- norbornyl p-bromobenzeneaulphonate (72). Ion-pair return would, of come, remain undetected in this way in the aolvolyais of the latter, and also in the aolvolyah of S,S-dirnethyl-~-norbornyl p-bromobenzenesulphonate (78) because of the eymmetry of the ion. All three of them Compounds underwent solvolyeie much more rapidly than the corresponding eado-isomera (Table 4) ;

Table 4. "he =tea of soetolyeis of some gem-dimethyl substituted norbornyl pbromobenzeneaulphonatea at 26'.

10% (eec-1) 10% (6ec-1) Norbornyl

p bromobenzenesulphonate ezo-isomer endo-isomer

7,7-Dimethyl 770 0.188 3,3-Dimethyl 32.9 0.0206 5,6-Dimethyl 20.1 0.039 Uneubstituted 88.2

. /

and the 7,7- and 3,3-dimethyl compounds, which. would be interconverted by a Wagner-Meerwein ehift, yielded almoet identical products (presumably isolated a t 100% reaction) on solvolysie in three solvents (Table 6). These are exclusively em-products, and no &product ( < 0.6%) W&B observed even when the gem-dimethyl group was a t position 7.48 Unfortunately, it ie

A. Colter, E. C. Friedriah, N. J. Holneae, and 5. Winetain, J . Am. Chsm. Soe., 87, 378

'I For the eignificsnce of this point BBB p. 12. (1886).

Clmsiccrl a )Id Non-clansical Garboniuin Ions 19

Table 5. Producta from the solvolyses of gem-dimebhylnorbomyl p - bromobenzeneaulphomtes.

Norbornyl p-bromobenzene-

sulphonate

At 25'

7,7-Me2 exo- (71) 3,3-Me2 exo- (72)

At 75" 7,7-Me2 em- (71) 7,7-Me2 endo- (74)

7,7-Me2 exn- (71) 3,3-Me2 em- (72) 5,5-Me2 exo- (78)

7,7-Me2 exo- (71) 7,7-Me2 endo- (74) 3,3-Me2 exo- (72) 3,3-Me2 endo- (75)

7,7-Dimethyl eZI0

47.0 47.0

43.0 43.0

(70.9)a 71.5 12.5

64.0 63.5 62.5 59.0

3,3-Dimethyl ex0

AcOH, 0.049m-NaOAc

4.5 4.0

0.0 6.0

72.4% Aqueous dioxan at 26'

(1 1.3)a 10.5 2.0

70% Aqueous acetone at 75"

11.5 10.5 12.0 13.5

a Infrared analysis: others by gas-liquid chromatography.

5,5-Dimethyl e20

48.5 49.0

51.0 51.0

(18.6)" 18.0 85.5

24.5 25.0 25.5 27.5

not clear if the small differences in ratios for the 7,7- and 3,3-dimethyl compounds in 70% aqueous acetone are significant. The 5,5-dimethyl system is interconverted with the other two through a 6 - t 2 (or 61 .1 ) hydride shift and the observation that there is less of this product in the more nucleophilic aqueous solvents than in acetic acid led the authors to conclude that this occm subeequent to, rather than concurrently with, the Wagner-Meerwein shift. It was also suggested that on theoreticaI grounds the 6 + 2 hydride shift probably involves an edge-protonated rather than a

(77) (79) (78)

face-protonated transition state (cf. p. 22). The endo-p-bromobenzenesulphonates (74) and (75) yield products in a very similar ratio to those from their em-isomers, a result which was interpreted as indicating efficient "leakage" from the classical ions (77) and (78) to the non-classical ion (79).

20 OrgCMIio Readion Mwha&ma 1965

Winatein and hie co-workers have also forcibly restated the case for re- gar- em-norbornyl and substituted em-norbornyl cations as having non- clagsical structure^.*^*^^

The products of the solvolysis of em-norbornyl p-bromobenzenesulphonate have been re-investigated and the high stereospecijioity of the reaction has been even more strikingly demonstrated.61 It is claimed that in 76% aqueous acetone the product of substitution contains less than 0.02% of d o - norbornanol, and that the m-norbornanol obtained from optically active p-bromobenzenesulphonate retains less than 0.06% of the optical activity.

Ethanolysia of camphene hydrochloride (80) and isobornyl chloride (81) has been shown to give mainly camphene (82) and the tertiary ether (82) with small amounts of the secondary ether (84) and tricyclene (85).61 It is not clear though whether the slightly smaller proportion of secondary ether formed from the tertiary camphene hydrochloride (see Table 6) is significant or not.

Ion-pair return occurs concurrently with acetolysis of 0-em-norbornyl p- trifluoromethylthiobenzoate (86) to yield the thiol ester (88).63 Positions 2 and 1 of the norbornyl system were labelled by using the optically active thiobenzoate (M), so that if the thiobenzoate ion returned to trap a classical norbornyl cation (87), the resulting thiobenzoate (88) would alao be optically active. Within experimental error ( > 97%), however, it was racemic, showing that the classical ion, if formed, rearrangea to its enantiomer (by movement ofthe l,&bond) more rapidly than it is trapped. The thiobenzoate ion is not,

R. Howe, E. C. Friedrioh, and 5. Winnbin, J . Am. Chcm. r9bc., 87,379 (1906). S. Winstaib, J . Am. Chen. ~ o c . , 87,381 (1906).

61 8. Winebin, E. Clippinger, R. Howe, and E. Vogelfanger, J . Am. Chem. Soc., 87,370 (1966). 6a C. A. Bunton end C . O’cOnnor, Chm. I d . (London), 1966, 1182. .w 8. Q. 8mith and J. P. Petroviah, J . Org. Chcm., 80,2882 (1906).

organic reaction mechanisms 1965€¦ · organic reaction mechanism generally and fairly uniformly,...

Documents